Blastp And Tblastx Bitscores

Question

[Damian Kao]

I am using BLAST 2.2.25+ command line version with default penalties.

I have a set of nucleotide sequences. I manually did a 6 frame translation with BioPython and blastp the 6 frames to NR protein.

Using the same set of nucleotide sequences, I also did a tblastx to another set of nucleotide sequences that is not in NR.

Would the bitscores from these two blasts be comparable?

Does tblastx essentially do a 6 frame translation on both nucleotide datasets and blastp them together?

Answer 1

[Michael Schubert]

Would the bitscores from these two blasts be comparable?

blastp and tblastx both score protein alignments. If you use the same substitution matrix, then yes, they should be comparable (even identical on the same set of query/subject sequences, but there might be some BLAST-internal optimization that yields slight differences).

Does tblastx essentially do a 6 frame translation on both nucleotide datasets and blastp them together?

Yes, that's essentially what it does. I.e., you don't need to do the translations by hand as long as you use a codon table supported by BLAST.

[Damian Kao]

Thanks! I was doing some ORF analysis, so I had the 6 frame translation done already. Figured I might as well save some time by just blastp instead of blastx.